RedPitayaDAQ

Python interfaces for collecting channel 1 voltage data on a Red Pitaya and launching acquisitions from a host computer over SSH. The installable package contains only src/RedPitayaDAQ.

Hardware acquisition has not been validated in this environment.

Installation

From the repository root, with Python 3.9 or newer:

python -m pip install .

For editable development, use python -m pip install -e .. NumPy, h5py, Matplotlib, and SciPy are declared dependencies. The package also imports the external recon_algorithms module, which must be available on PYTHONPATH; it is not yet supplied by this package. The local driver additionally requires the Red Pitaya vendor rp Python bindings, supplied by the device environment; they are not installed from PyPI by this package. The remote interface uses system ssh and scp executables and requires working SSH authentication.

from RedPitayaDAQ import RedPitayaRemoteInterface

remote = RedPitayaRemoteInterface(ip="192.168.137.111", local_dir="rp_data")
remote.update_laser_params(power_mW=3.0, wavelength_nm=1550)

The package root does not import rp. On the device, import the driver explicitly:

from RedPitayaDAQ.RedPitayaLocalDriver import RedPitayaLocalDriver

driver = RedPitayaLocalDriver()

Local acquisition API

RedPitayaLocalDriver runs on the device. Channel 1's physical jumper must be set to the high-voltage range. Its methods are:

Method	Purpose
check(result, operation)	Raise RuntimeError when an API status differs from rp.RP_OK.
setParameters(sampling_rate_wanted, scan_duration, chunk_samples)	Derive decimation and rate from a 125 MHz clock, align chunks down to eight samples, and calculate sample/chunk counts.
collectData(sampling_rate_wanted=32000, scan_duration=10, chunk_samples=32000, start_timestamp=None, extra_metadata=None)	Acquire voltage chunks, write HDF5 files, and return a relative Path to the run directory.

Sampling rates are in samples per second and durations in seconds. The actual rate is 125000000 / round(125000000 / requested_rate). Inputs are not validated; use positive rates and durations and chunk sizes of at least eight samples within the device's buffer capacity. The final chunk is collected in full, so the run may contain more samples than requested. Acquisition occurs in separate chunks, with setup and file-writing time between them.

start_timestamp uses %Y%m%d_%H%M%S; when omitted, the device's local time is used. Data is written relative to the working directory:

<timestamp>_N<total_chunks>/
    <timestamp>_chunk1_of<total_chunks>.h5
    ...

Each file contains a one-dimensional Voltage (V) dataset and file attributes: date, total_samples, chunk_samples, total_chunks, current_chunk, scan_duration, sampling_rate, and decimation. Additional metadata accepts scalar numbers and strings. Hardware stop, AXI disable, and release are attempted when leaving the acquisition block.

Each chunk has a buffer-fill timeout equal to its expected sampling duration plus 10 seconds. A timeout raises TimeoutError and runs the hardware cleanup block. Buffer status is polled with a 1 ms sleep between unsuccessful checks.

Remote acquisition API

RedPitayaRemoteInterface(ip="192.168.137.111", local_dir="rp_data") defaults to SSH user root and device working directory /home/jupyter/RedPitaya. These are stored in RP_USER and RP_WORKDIR; the latter is a PurePosixPath.

Method	Purpose
update_laser_params(power_mW=None, wavelength_nm=None)	Set metadata values; does not control a laser.
update_local_datadir(local_folder)	Change the download destination.
run_acquisition(sampling_rate_sps=None, scan_duration_sec=None)	Launch the device command, stream stdout, and return (remote_folder, elapsed_seconds). Supplied overrides persist on the instance.
copy_last_acquisition(remote_output_folder, local_folder_override=None, remove=True)	Recursively copy a run using SCP and return its local Path. By default, delete the remote run after successful copying.

Defaults are 32,000 samples/s, 900 seconds, 960,000 samples per chunk, and laser metadata of 3 mW and 1550 nm. Elapsed time includes the SSH command's startup and execution overhead. Supply trusted run directory names to the copy method and use remove=False to retain data on the device.

The remote command imports RedPitayaDAQ.RedPitayaLocalDriver.RedPitayaLocalDriver. Install the package in the device's python3 environment and ensure its dependencies, including the external recon_algorithms module, are available. The command adds /opt/redpitaya/lib/python to PYTHONPATH for the vendor bindings.

Command-line interface

The remote module can be invoked with:

python -m RedPitayaDAQ --help

The remote module provides parse_cli_args() and run_full_daq(); the latter starts acquisition and downloads with remote deletion.

Option	Short	Default	Validation
--ip	-i	192.168.137.111	IPv4 address
--outdir	-o	rp_data	Local destination
--samprate	-r	32000	Integer in (0, 100000] samples/s
--duration	-d	900	(0, 3600] seconds
--laserpower	-l	3.0	[0, 17] mW
--laserwavelen	-w	1550	[1520, 1570] nm

Checked invalid values produce messages and fall back to defaults. These CLI checks do not apply to direct method calls.

Repository layout

src/RedPitayaDAQ contains the packaged interfaces. legacy/ holds earlier laptop/device control and runner scripts. tools/ contains reconstruction algorithms and a data-checking notebook. Neither directory is installed as part of the package.

Function and method docstrings added in this documentation pass carry [AI generated]; revised existing method docstrings carry [AI edit].